Global and regional left ventricular myocardial deformation measures by magnetic resonance feature tracking in healthy volunteers: comparison with tagging and relevance of gender

Abstract

Background: Feature tracking software offers measurements of myocardial strain, velocities and displacement from cine cardiovascular magnetic resonance (CMR) images. We used it to record deformation parameters in healthy adults and compared values to those obtained by tagging.

Methods: We used TomTec 2D Cardiac Performance Analysis software to derive global, regional and segmental myocardial deformation parameters in 145 healthy volunteers who had steady state free precession (SSFP) cine left ventricular short (basal, mid and apical levels) and long axis views (horizontal long axis, vertical long axis and left ventricular out flow tract) obtained on a 1.5 T Siemens Sonata scanner. 20 subjects also had tagged acquisitions and we compared global and regional deformation values obtained from these with those from Feature Tracking.

Results: For globally averaged measurements of strain, only those measured circumferentially in short axis slices showed reasonably good levels of agreement between FT and tagging (limits of agreement -0.06 to 0.04). Longitudinal strain showed wide limits of agreement (-0.16 to 0.03) with evidence of overestimation of strain by FT relative to tagging as the mean of both measures increased. Radial strain was systematically overestimated by FT relative to tagging with very wide limits of agreement extending to as much as 100% of the mean value (-0.01 to 0.23). Reproducibility showed similar relative trends with acceptable global inter-observer variability for circumferential measures (coefficient of variation 4.9%) but poor reproducibility in the radial direction (coefficient of variation 32.3%). Ranges for deformation parameters varied between basal, mid and apical LV levels with higher levels at base compared to apex, and between genders by both FT and tagging.

Conclusions: FT measurements of circumferential but not longitudinally or radially directed global strain showed reasonable agreement with tagging and acceptable inter-observer reproducibility. We record provisional ranges of FT deformation parameters at global, regional and segmental levels. They show evidence of variation with gender and myocardial region in the volunteers studied, but have yet to be compared with tagging measurements at the segmental level.

Figure 1

Modified Bland Altman plots showing comparisons between Feature Tracking and CMR tagging for global strain parameters: Circumferential strain (left); longitudinal strain (middle); radial strain (right). The bias (blue solid line) and limits of agreement (blue dashed lines) are shown. The oblique dashed lines demonstrate 25 (green), 50 (purple) and 100% (red) difference levels.